US plans broader controls on overseas access to AI computing
The Trump administration is preparing interim export controls intended to stop companies, including Chinese firms, from reaching U.S. AI computing capacity through data centers in other countries. The Commerce Department would require licenses to export, reexport or transfer AI chips and servers to facilities worldwide when they could provide computing power to companies in China or other restricted countries.
The rules target the D:5 group, which includes China, Russia, Iran and North Korea, and would also cover indirect access arranged through brokers or other intermediaries. Operators and exporters would have to examine customers’ ownership and activities. The requirements apply to networks capable of more than 12 million total processing performance, roughly the capacity of 750 Nvidia H100 or H200 chips. Existing operators would have 90 days to obtain licenses.
The proposal adds licensing requirements for advanced chips sent to Malaysia, Singapore, Indonesia, Thailand and Brazil, though shipments intended for U.S.-based companies would be exempt. It excludes commercial recommendation systems. The measure would also withdraw the Biden administration’s January 2025 AI Diffusion Rule, which imposed country-based limits on advanced AI chip exports and certain model transfers. The move follows reports that Chinese AI companies are seeking overseas computing capacity and that Megaspeed allegedly obtained Nvidia chips for Chinese customers.
Written by Sakhanda Wire AI from reporting by TechCrunch. Facts, quotes and full context are in the original article below.